Fourth Annual PRIMES Conference . MAY 18, 2014
Loop Extruding
Enzymes in Interphase: Dynamic
Folding of Chromatin Domains
Carolyn Lu
Professor Leonid Mirny
Maxim Imakaev, Geoffrey Fudenberg
Characterizing Chromosomal Contacts:
Contact maps and scalings
Contact maps
Contact probability scalings:
(contact probability vs. separation distance)
contacts
many
contacts
distance
few
contacts
TADs: Topologically Associated Domains
55.5 Mb
few
contacts
many
contacts
● Regions of increased
interaction
● Decreased interaction
across boundaries
6.5 Mb
~1.3Mb
62 Mb
55.5 Mb
Region of Chromosome 15
62 Mb
contact frequency
TAD Scalings
● Within TAD
slope in log-log
of contact
probability over
distance in kb is
-0.5
● Between TADs
slope goes from
around -0.5 at
closer distances
to -1.0 at further
distance in kb
What mechanisms can define TADs
and TAD boundaries consistent with
experimental data?
Methods: Polymer simulations
Experimental
Simulation
?
?
contact
map
scaling?
??
Previously explored models:
thick-thin
stiff-flexible
RNA
SMCs as loop-extruding proteins
Alipour, Elnaz, and John F.
Marko. "Self-organization of
domain structures by DNA-loopextruding enzymes." Nucleic
acids research 40.22 (2012):
11202-11212.
Anton Goloborodko & Leonid Mirny
(Unpublished)
Basic kinetic model of SMC Loop
Extrusion
SMC binds DNA
SMC extrudes DNA
SMC releases DNA
SMC loop extrusion + boundary
→ TAD boundaries?
TAD boundaries have been shown to be highly bound by a
number of key epigenetic regulators.
Hypothesis: Boundaries which halt SMC loop extrusion or
release SMC loops can in turn create TAD boundaries
boundary
Complications of the kinetic model
1. What happens when a loop meets a
boundary?
2. What happens when a loop meets another
loop?
Kinetic SMC model: Loop-boundary behavior
Boundary
OR
Stall
Release
Kinetic SMC model: loop-loop behavior
Stall
OR
Cross
Four variations on the basic kinetic
model of SMC loop extrusion
Since details of SMC loop extrusion are unknown, we
tested a number of possible variations:
- Boundary stalling and loop-loop stalling (stalling/stalling)
- Long boundary stalling and loop-loop stalling (long-stalling/stalling)
- Boundary release and loop-loop crossing (release/crossing)
- Boundary release and loop-loop stalling (release/stalling)
SMC extrusion kinetic Model 1:
Boundary stalling and loop-loop stalling
time
monomer
boundary
SMC extrusion kinetic model 1:
Contact Map (stall/stall)
Simulated TADs
Experimental TADs
contact frequency
Stall-stall: Scalings
SMC extrusion kinetic Model 2: Long
boundaries and loop-loop stalling
monomer
time
Point boundary stall/stall
Long boundary: Contact maps
Loop density 20
Loop density 5
Experimental TADs
Long boundary: Scalings
Loop density 20
Loop density 5
SMC extrusion Model 3: Boundary
release and loop crossing
monomer
time
Point boundary stall/stall
Release/cross: Contact map
Simulated TADs
Experimental TADs
contact frequency
Release/cross: Scalings
monomer distance
SMC extrusion Model 4: Boundary
release and loop stalling
Point boundary stall/stall
monomer
time
Model 3: release/cross
Release-stall: Contact map
Simulated TADs
Experimental TADs
contact frequency
Release-stall: Scalings
monomer distance
experimental
contact frequency
Comparison of
models
monomer distance
Conclusions and Discussion
● Our models of SMC loop extrusion can reproduce some features of TAD
contact maps but do not completely reproduce TAD scalings
● Details of SMC behavior strongly affect simulated contact maps and
scalings
o Release at boundaries / stalling between loops best reproduced
experimental data
● Further biological experiments are necessary to determine detailed
mechanisms of SMC action & TAD formation
Thanks to PRIMES alum/MIT
student Boryana Doyle, awesome
mentors Geoffrey Fudenberg and
Maxim Imakaev, Prof. Leonid
Mirny, and MIT PRIMES.